Continuing the previous problem, analyze part b in Excel as follows. Start with hypothetical population mean GMAT scores for men and women, along with population standard deviations. Enter these at the top of a worksheet. You can make the two means as close as you like, but not identical. Simulate a sample of men’s GMAT scores with your mean and standard deviation in column A. Do the same for women in column B. The sample sizes do not have to be the same, but you can make them the same. Then run the test for the difference between two means. (The point of this problem is that if the population means are fairly close and you pick relatively small sample sizes, the sample mean differences probably won’t be significant.
If you find this, generate new samples of a larger sample size and redo the test. Now they might be significant. If not, try again with a still larger sample size. Eventually, you should get statistically significant differences.)